LAFS: Landmark-based Facial Self-supervised Learning for Face Recognition
Open paper
Benchmark
Model
Rank
Results
face-identification-on-megaface
Part fViT, ours (LAFS)
#7
Accuracy: 98.66%
face-verification-on-ijb-c
Part fViT, ours (LAFS)
#28
TAR @ FAR=1e-4: 97.48%
Rank counts only results with a code link.